A defense for the PhD proposal of Raul Penagui´...
Transcript of A defense for the PhD proposal of Raul Penagui´...
![Page 1: A defense for the PhD proposal of Raul Penagui´ ao˜user.math.uzh.ch/penaguiao/docs/Slides/defense_beamer-Handout.pdf · Introduction Patterns in combinatorics Graph invariants Generalized](https://reader033.fdocuments.us/reader033/viewer/2022060508/5f2381e3f00bc922b015f1ad/html5/thumbnails/1.jpg)
Introduction Patterns in combinatorics Graph invariants Generalized permutahedra Future work from my thesis
Hopf algebras and combinatoricsA defense for the PhD proposal of
Raul Penaguiao
University of Zurich
28th May, 2020
Slides can be found inhttp://user.math.uzh.ch/penaguiao/
![Page 2: A defense for the PhD proposal of Raul Penagui´ ao˜user.math.uzh.ch/penaguiao/docs/Slides/defense_beamer-Handout.pdf · Introduction Patterns in combinatorics Graph invariants Generalized](https://reader033.fdocuments.us/reader033/viewer/2022060508/5f2381e3f00bc922b015f1ad/html5/thumbnails/2.jpg)
Introduction Patterns in combinatorics Graph invariants Generalized permutahedra Future work from my thesis
What is algebraic combinatorics - combinatorics
What are combinatorial objects?
1342
Associated with a notion of not only a way to count, but also ofsize and how to merged and split them.
![Page 3: A defense for the PhD proposal of Raul Penagui´ ao˜user.math.uzh.ch/penaguiao/docs/Slides/defense_beamer-Handout.pdf · Introduction Patterns in combinatorics Graph invariants Generalized](https://reader033.fdocuments.us/reader033/viewer/2022060508/5f2381e3f00bc922b015f1ad/html5/thumbnails/3.jpg)
Introduction Patterns in combinatorics Graph invariants Generalized permutahedra Future work from my thesis
What is algebraic combinatorics - algebra
Goal: associate to combinatorial objects some algebraicstrucutre (Groups, rings, algebras, monoidal categories).Special algebraic structure: Hopf algebras (H,µ,∆, ι, ε, S)
• Start with an algebra (H,µ, ι).• Add a coproduct ∆ with a counit ε.• Add an antipode S : H → H, that plays a role of “ inverse
function ”.
Example of a Hopf algebra: the usual algebra K[x] with thecoproduct
∆ : x 7→ x⊗ 1 + 1⊗ x
S : xn 7→ (−x)n .
![Page 4: A defense for the PhD proposal of Raul Penagui´ ao˜user.math.uzh.ch/penaguiao/docs/Slides/defense_beamer-Handout.pdf · Introduction Patterns in combinatorics Graph invariants Generalized](https://reader033.fdocuments.us/reader033/viewer/2022060508/5f2381e3f00bc922b015f1ad/html5/thumbnails/4.jpg)
Introduction Patterns in combinatorics Graph invariants Generalized permutahedra Future work from my thesis
Main goals in Hopf algebras in combinatorics
What types of problems do we want to solve in algebraiccombinatorics? Simple formulas for the antipode, structuretheorems (freeness, cofreeness) and chromatic invariants.
![Page 5: A defense for the PhD proposal of Raul Penagui´ ao˜user.math.uzh.ch/penaguiao/docs/Slides/defense_beamer-Handout.pdf · Introduction Patterns in combinatorics Graph invariants Generalized](https://reader033.fdocuments.us/reader033/viewer/2022060508/5f2381e3f00bc922b015f1ad/html5/thumbnails/5.jpg)
Introduction Patterns in combinatorics Graph invariants Generalized permutahedra Future work from my thesis
Introduction
Patterns in combinatorics
Graph invariants
Generalized permutahedra
Future work from my thesis
![Page 6: A defense for the PhD proposal of Raul Penagui´ ao˜user.math.uzh.ch/penaguiao/docs/Slides/defense_beamer-Handout.pdf · Introduction Patterns in combinatorics Graph invariants Generalized](https://reader033.fdocuments.us/reader033/viewer/2022060508/5f2381e3f00bc922b015f1ad/html5/thumbnails/6.jpg)
Introduction Patterns in combinatorics Graph invariants Generalized permutahedra Future work from my thesis
Permutations and patternsA permutation π of size n is an arrangement on an n× n table:
π = = 2431
The set of permutations of size n : Sn. The set of allpermutations : S.Select a set I of columns of the square configuration of π anddefine the restriction π|I . This is a permutation.
π|{1,2,4} = 231=
![Page 7: A defense for the PhD proposal of Raul Penagui´ ao˜user.math.uzh.ch/penaguiao/docs/Slides/defense_beamer-Handout.pdf · Introduction Patterns in combinatorics Graph invariants Generalized](https://reader033.fdocuments.us/reader033/viewer/2022060508/5f2381e3f00bc922b015f1ad/html5/thumbnails/7.jpg)
Introduction Patterns in combinatorics Graph invariants Generalized permutahedra Future work from my thesis
Pattern functions
We can count occurrences!For permutations π, σ, we define the pattern number:
pπ(σ) = #{occurrences of π as a restriction of σ} .
In this way we have
p12(4132) = 2, p312(4132) = 2, p12(12345) = 10
and p312(3675421) = 0
The same objects can be defined on partitions and graphs andmany more! Pattern functions come from combinatorialpresheaves.
![Page 8: A defense for the PhD proposal of Raul Penagui´ ao˜user.math.uzh.ch/penaguiao/docs/Slides/defense_beamer-Handout.pdf · Introduction Patterns in combinatorics Graph invariants Generalized](https://reader033.fdocuments.us/reader033/viewer/2022060508/5f2381e3f00bc922b015f1ad/html5/thumbnails/8.jpg)
Introduction Patterns in combinatorics Graph invariants Generalized permutahedra Future work from my thesis
Pattern Hopf algebras
Fact: the pattern functions {pπ}π∈S form a linearly independentset.
A(Per) := span{pπ}π∈S .
Theorem (Vargas, 2014)The space A(Per) is a Hopf algebra that is freely generated bythe so called Lyndon permutations.
Theorem (P., 2019)Let h be an associative presheaf. The space A(h) is a Hopfalgebra. This is a free algebra when h is a commutativepresheaf or is the presheaf on marked permutations.
![Page 9: A defense for the PhD proposal of Raul Penagui´ ao˜user.math.uzh.ch/penaguiao/docs/Slides/defense_beamer-Handout.pdf · Introduction Patterns in combinatorics Graph invariants Generalized](https://reader033.fdocuments.us/reader033/viewer/2022060508/5f2381e3f00bc922b015f1ad/html5/thumbnails/9.jpg)
Introduction Patterns in combinatorics Graph invariants Generalized permutahedra Future work from my thesis
Meanwhile in Twitter
![Page 10: A defense for the PhD proposal of Raul Penagui´ ao˜user.math.uzh.ch/penaguiao/docs/Slides/defense_beamer-Handout.pdf · Introduction Patterns in combinatorics Graph invariants Generalized](https://reader033.fdocuments.us/reader033/viewer/2022060508/5f2381e3f00bc922b015f1ad/html5/thumbnails/10.jpg)
Introduction Patterns in combinatorics Graph invariants Generalized permutahedra Future work from my thesis
Feasible regions
Fix Sj =⋃j≥kSk of permutations. What are the possible values
of(pπ(σ) |σ|−|π|
)π∈Sj
∈ RSj when |σ| is big? These are the so
called feasible values and form the feasible region.
Theorem (R. Glebov, C. Hoppen, et al, 2017)The feasible region has dimension at least
|{ indecomposable permutations of size ≤ j }| .
Conjecture (J. Borga, P.)The dimension of the feasible region for classical occurrencesis enumerated by the Lyndon permutations of size at most j.
![Page 11: A defense for the PhD proposal of Raul Penagui´ ao˜user.math.uzh.ch/penaguiao/docs/Slides/defense_beamer-Handout.pdf · Introduction Patterns in combinatorics Graph invariants Generalized](https://reader033.fdocuments.us/reader033/viewer/2022060508/5f2381e3f00bc922b015f1ad/html5/thumbnails/11.jpg)
Introduction Patterns in combinatorics Graph invariants Generalized permutahedra Future work from my thesis
Consecutive patterns
We now consider only occurrences that form an interval. Forinstance, taking σ = 2413, there are two distinct consecutiverestructions of σ of size three, namely 231 and 312.
Theorem (J. Borga, P., 2019)The feasible region for consecutive patterns is a polytope.Specifically, it is the cycle polytope of overlap graphs onpermutations.
![Page 12: A defense for the PhD proposal of Raul Penagui´ ao˜user.math.uzh.ch/penaguiao/docs/Slides/defense_beamer-Handout.pdf · Introduction Patterns in combinatorics Graph invariants Generalized](https://reader033.fdocuments.us/reader033/viewer/2022060508/5f2381e3f00bc922b015f1ad/html5/thumbnails/12.jpg)
Introduction Patterns in combinatorics Graph invariants Generalized permutahedra Future work from my thesis
Meanwhile in Twitterland
![Page 13: A defense for the PhD proposal of Raul Penagui´ ao˜user.math.uzh.ch/penaguiao/docs/Slides/defense_beamer-Handout.pdf · Introduction Patterns in combinatorics Graph invariants Generalized](https://reader033.fdocuments.us/reader033/viewer/2022060508/5f2381e3f00bc922b015f1ad/html5/thumbnails/13.jpg)
Introduction Patterns in combinatorics Graph invariants Generalized permutahedra Future work from my thesis
What is a graph
A graph is a pair of sets (V,E). For instance:
A partition is a list of positive integers (λ1 ≤ · · · ≤ λk). Fixed apartition λ = (λ1 ≤ · · · ≤ λk), let Kc
λ be a graph:Kcλ = (V1 ∪ · · · ∪ Vk, E) with maximal edges in such a way that
each Vi is independent.
Figure: The graph Kc(1,2,3).
![Page 14: A defense for the PhD proposal of Raul Penagui´ ao˜user.math.uzh.ch/penaguiao/docs/Slides/defense_beamer-Handout.pdf · Introduction Patterns in combinatorics Graph invariants Generalized](https://reader033.fdocuments.us/reader033/viewer/2022060508/5f2381e3f00bc922b015f1ad/html5/thumbnails/14.jpg)
Introduction Patterns in combinatorics Graph invariants Generalized permutahedra Future work from my thesis
Another Hopf algebra structure on graphsProduct = disjoint union of graphs.
Figure: Product structure on the graph Hopf algebra.
⊗
Figure: Coproduct structure on the graph Hopf algebra.
This defines mapsµ : G⊗G→ G
∆ : G→ G⊗G
![Page 15: A defense for the PhD proposal of Raul Penagui´ ao˜user.math.uzh.ch/penaguiao/docs/Slides/defense_beamer-Handout.pdf · Introduction Patterns in combinatorics Graph invariants Generalized](https://reader033.fdocuments.us/reader033/viewer/2022060508/5f2381e3f00bc922b015f1ad/html5/thumbnails/15.jpg)
Introduction Patterns in combinatorics Graph invariants Generalized permutahedra Future work from my thesis
Chromatic numbersGiven a graph H = (V,E), a stable coloring f : V → N is suchthat two neighboring vertices have always distinct colors.
1
1
2
2
4
Figure: A stable coloring of a graph.
Chromatic number = lowest number of colors that still allows fora stable colouring. NP complete problem (one of the originalones).
χH(n) = #{ graph-colorings with n colors } .
![Page 16: A defense for the PhD proposal of Raul Penagui´ ao˜user.math.uzh.ch/penaguiao/docs/Slides/defense_beamer-Handout.pdf · Introduction Patterns in combinatorics Graph invariants Generalized](https://reader033.fdocuments.us/reader033/viewer/2022060508/5f2381e3f00bc922b015f1ad/html5/thumbnails/16.jpg)
Introduction Patterns in combinatorics Graph invariants Generalized permutahedra Future work from my thesis
Deletion - contraction relations
Given a graph H, we can define the deletion and acontraction of a set of edges.
H
e
H \ {e} H/{e}H
e
Figure: Deletion and contraction of edges.
χH(n) = χH\e(n)− χH/e(n) .
![Page 17: A defense for the PhD proposal of Raul Penagui´ ao˜user.math.uzh.ch/penaguiao/docs/Slides/defense_beamer-Handout.pdf · Introduction Patterns in combinatorics Graph invariants Generalized](https://reader033.fdocuments.us/reader033/viewer/2022060508/5f2381e3f00bc922b015f1ad/html5/thumbnails/17.jpg)
Introduction Patterns in combinatorics Graph invariants Generalized permutahedra Future work from my thesis
Reciprocity results
Given a graph, H, what is the meaning of χH(−n)?
Theorem (R. Stanley, 1973)
χH(−1) = (−1)n#{ acyclic orientations of the graph H } .
Interpretations of χH(−k) for k > 1 also exist that relate thechromatic polynomial and acyclic orientations.
![Page 18: A defense for the PhD proposal of Raul Penagui´ ao˜user.math.uzh.ch/penaguiao/docs/Slides/defense_beamer-Handout.pdf · Introduction Patterns in combinatorics Graph invariants Generalized](https://reader033.fdocuments.us/reader033/viewer/2022060508/5f2381e3f00bc922b015f1ad/html5/thumbnails/18.jpg)
Introduction Patterns in combinatorics Graph invariants Generalized permutahedra Future work from my thesis
The antipode
Theorem (B. Humpert and J. Martin, 2012)Let H = (V,E) be a graph, then
S(H) =∑
F⊆E flat
(−1)n−rank(F )a(H/F )H(V,F ) .
G
G
K[x]
K[x]
S S
χ
χ
χH(−n) = S(χH(n)) = χS(H)(n) .
![Page 19: A defense for the PhD proposal of Raul Penagui´ ao˜user.math.uzh.ch/penaguiao/docs/Slides/defense_beamer-Handout.pdf · Introduction Patterns in combinatorics Graph invariants Generalized](https://reader033.fdocuments.us/reader033/viewer/2022060508/5f2381e3f00bc922b015f1ad/html5/thumbnails/19.jpg)
Introduction Patterns in combinatorics Graph invariants Generalized permutahedra Future work from my thesis
The ring of symmetric functions
A symmetric function on the variables x1, . . . (infinitely manyvariables) is a formal sum of monomials on x1, . . . withbounded degree.Examples: 1 , x1 + x2 + . . . , x21 + x22 + . . .
x1x22 + x21x2 + x1x
23 + x21x3 + x2x
23 + x22x3 + x1x
24 . . .
This forms a Hopf algebra.
![Page 20: A defense for the PhD proposal of Raul Penagui´ ao˜user.math.uzh.ch/penaguiao/docs/Slides/defense_beamer-Handout.pdf · Introduction Patterns in combinatorics Graph invariants Generalized](https://reader033.fdocuments.us/reader033/viewer/2022060508/5f2381e3f00bc922b015f1ad/html5/thumbnails/20.jpg)
Introduction Patterns in combinatorics Graph invariants Generalized permutahedra Future work from my thesis
The CSF
The ultimate chromatic invariant.
ΨG(H) =∑
f stable coloring
xf(1) . . . xf(n) =∑
f stable coloring
xf ∈ Sym .
Examples of CSF:
ΨG
( )= 6(x1x2x3 + x1x2x4 + x1x3x4 + . . . ) ,
ΨG
( )= 4(x21x2x3 + . . . ) + 24(x1x2x3x4 + . . . ) .
![Page 21: A defense for the PhD proposal of Raul Penagui´ ao˜user.math.uzh.ch/penaguiao/docs/Slides/defense_beamer-Handout.pdf · Introduction Patterns in combinatorics Graph invariants Generalized](https://reader033.fdocuments.us/reader033/viewer/2022060508/5f2381e3f00bc922b015f1ad/html5/thumbnails/21.jpg)
Introduction Patterns in combinatorics Graph invariants Generalized permutahedra Future work from my thesis
The tree conjecture
Conjecture (Tree conjecture)Given two non-isomorphic trees T1, T2, their chromaticsymmetric functions are distinct, i.e. ΨG(T1) 6= ΨG(T2).
Partial results: Proper caterpillars, “ labelled ” case.
Figure: A tree that is a proper caterpillar.
![Page 22: A defense for the PhD proposal of Raul Penagui´ ao˜user.math.uzh.ch/penaguiao/docs/Slides/defense_beamer-Handout.pdf · Introduction Patterns in combinatorics Graph invariants Generalized](https://reader033.fdocuments.us/reader033/viewer/2022060508/5f2381e3f00bc922b015f1ad/html5/thumbnails/22.jpg)
Introduction Patterns in combinatorics Graph invariants Generalized permutahedra Future work from my thesis
Modular relations on graphs
Are there any linear relations that ΨG satisfy?
Figure: Modular relations that annihilate ΨG, [GP13] and [OM14].
Figure: Two unicyclic graphs with the same CSF.
![Page 23: A defense for the PhD proposal of Raul Penagui´ ao˜user.math.uzh.ch/penaguiao/docs/Slides/defense_beamer-Handout.pdf · Introduction Patterns in combinatorics Graph invariants Generalized](https://reader033.fdocuments.us/reader033/viewer/2022060508/5f2381e3f00bc922b015f1ad/html5/thumbnails/23.jpg)
Introduction Patterns in combinatorics Graph invariants Generalized permutahedra Future work from my thesis
The kernel problem
Theorem (P., 2018)The kernel of ΨG : G→ Sym is generated by modularrelations.The tree conjecture is established if we find a graph invariant ζthat satisfies the following two properties• The invariant ζ distinguishes trees, i.e. for two
non-isomorphic trees T1, T2 we have that ζ(T1) 6= ζ(T2).• The invariant ζ satisfies the modular relations.
![Page 24: A defense for the PhD proposal of Raul Penagui´ ao˜user.math.uzh.ch/penaguiao/docs/Slides/defense_beamer-Handout.pdf · Introduction Patterns in combinatorics Graph invariants Generalized](https://reader033.fdocuments.us/reader033/viewer/2022060508/5f2381e3f00bc922b015f1ad/html5/thumbnails/24.jpg)
Introduction Patterns in combinatorics Graph invariants Generalized permutahedra Future work from my thesis
Meanwhile in Twitter
![Page 25: A defense for the PhD proposal of Raul Penagui´ ao˜user.math.uzh.ch/penaguiao/docs/Slides/defense_beamer-Handout.pdf · Introduction Patterns in combinatorics Graph invariants Generalized](https://reader033.fdocuments.us/reader033/viewer/2022060508/5f2381e3f00bc922b015f1ad/html5/thumbnails/25.jpg)
Introduction Patterns in combinatorics Graph invariants Generalized permutahedra Future work from my thesis
The kernel solutionProof: Start with an expression
∑i
αiGi ∈ ker ΨG. Always
increase the number of edges on a graph!Consider the path of length three:
+ − + −
−+
For which graphs aren’t we able to add edges? Graphs of theform Kc
λ. Fact: {ΨG(Kcλ)}λ`n is a linearly independent set.
![Page 26: A defense for the PhD proposal of Raul Penagui´ ao˜user.math.uzh.ch/penaguiao/docs/Slides/defense_beamer-Handout.pdf · Introduction Patterns in combinatorics Graph invariants Generalized](https://reader033.fdocuments.us/reader033/viewer/2022060508/5f2381e3f00bc922b015f1ad/html5/thumbnails/26.jpg)
Introduction Patterns in combinatorics Graph invariants Generalized permutahedra Future work from my thesis
Polytopes - SimplicesLet’s go to the n-dimentional space Rn, that has a canonicalbasis {~ei|i = 1, . . . , n}.
SimI = conv{~ei|i ∈ I} .
e3
e1 e2
A
BC
Figure: The polytopes A = Sim{1}, B = Sim{1,3}, C = Sim{1,2,3} aresimplices in R3.
![Page 27: A defense for the PhD proposal of Raul Penagui´ ao˜user.math.uzh.ch/penaguiao/docs/Slides/defense_beamer-Handout.pdf · Introduction Patterns in combinatorics Graph invariants Generalized](https://reader033.fdocuments.us/reader033/viewer/2022060508/5f2381e3f00bc922b015f1ad/html5/thumbnails/27.jpg)
Introduction Patterns in combinatorics Graph invariants Generalized permutahedra Future work from my thesis
Polytopes - Minkowski operationsGiven two convex sets A,B and a non-negative real number λ,their Minkowski sum is given by A+B = {a+ b|a ∈ A, b ∈ B}and the Minkowski dilation is λA = {λa|a ∈ A}.
+ =
For convex sets A,B, the set A−B is the convex set C suchthat B + C = A. May not exists, but it is unique.
![Page 28: A defense for the PhD proposal of Raul Penagui´ ao˜user.math.uzh.ch/penaguiao/docs/Slides/defense_beamer-Handout.pdf · Introduction Patterns in combinatorics Graph invariants Generalized](https://reader033.fdocuments.us/reader033/viewer/2022060508/5f2381e3f00bc922b015f1ad/html5/thumbnails/28.jpg)
Introduction Patterns in combinatorics Graph invariants Generalized permutahedra Future work from my thesis
Polytopes - Generalized permutahedraA generalized permutahedron is a polytope that results as(∑
aJ>0
aJ SimJ
)−
(∑aJ<0
|aJ | SimJ
).
A hypergraphic polytope is a polytope of the form∑aJ>0
aJ SimJ .
Figure: The permutahedra for n = 4.
![Page 29: A defense for the PhD proposal of Raul Penagui´ ao˜user.math.uzh.ch/penaguiao/docs/Slides/defense_beamer-Handout.pdf · Introduction Patterns in combinatorics Graph invariants Generalized](https://reader033.fdocuments.us/reader033/viewer/2022060508/5f2381e3f00bc922b015f1ad/html5/thumbnails/29.jpg)
Introduction Patterns in combinatorics Graph invariants Generalized permutahedra Future work from my thesis
Faces and linear optimization
Given a polytope q, a face of q is a set of the form
qf := minx∈q
f(x) ,
for some linear function f .
Figure: For some linear functionals the corresponding face.
![Page 30: A defense for the PhD proposal of Raul Penagui´ ao˜user.math.uzh.ch/penaguiao/docs/Slides/defense_beamer-Handout.pdf · Introduction Patterns in combinatorics Graph invariants Generalized](https://reader033.fdocuments.us/reader033/viewer/2022060508/5f2381e3f00bc922b015f1ad/html5/thumbnails/30.jpg)
Introduction Patterns in combinatorics Graph invariants Generalized permutahedra Future work from my thesis
A chromatic invariant on generalized permutahedra
xα11 xα2
2 · · · 7→ (α1, α2, . . . ) .
Define ΨGP(q) =∑
α vector
x(α)1[qα is a point ] .
This is a chromatic invariant as well! But is not in Sym anymore.
GP
G
QSym
Sym
ΨGP
ΨG
Z
Figure: Commutative diagram of chromatic invariants.
Theorem (P. , 2018)Explicit generators of the kernel of ΨGP can be given, whenrestricted to HGP.
![Page 31: A defense for the PhD proposal of Raul Penagui´ ao˜user.math.uzh.ch/penaguiao/docs/Slides/defense_beamer-Handout.pdf · Introduction Patterns in combinatorics Graph invariants Generalized](https://reader033.fdocuments.us/reader033/viewer/2022060508/5f2381e3f00bc922b015f1ad/html5/thumbnails/31.jpg)
Introduction Patterns in combinatorics Graph invariants Generalized permutahedra Future work from my thesis
Future work• Find all modular relations for the chromatic invariant onGP.• Freeness conjecture on pattern algebras, antipode
formulas for pattern algebras.• More exciting feasible regions!
(1,0,0,0,0,0)
(0, 12, 12, 0, 0, 0)
( 12, 14, 14, 0, 0, 0)
(0, 0, 12, 12, 0, 0)
( 13, 0, 1
3, 13, 0, 0)
(0, 0, 0, 12, 12, 0)
(0, 12, 0, 1
2, 0, 0)
( 13, 13, 0, 1
3, 0, 0)
Figure: The feasible region for consecutive occurrences ofpermutations avoiding 321.
![Page 32: A defense for the PhD proposal of Raul Penagui´ ao˜user.math.uzh.ch/penaguiao/docs/Slides/defense_beamer-Handout.pdf · Introduction Patterns in combinatorics Graph invariants Generalized](https://reader033.fdocuments.us/reader033/viewer/2022060508/5f2381e3f00bc922b015f1ad/html5/thumbnails/32.jpg)
Introduction Patterns in combinatorics Graph invariants Generalized permutahedra Future work from my thesis
The end